ValueError: Unable to create dataset (name already exists)

[Syed Shoaib Abbas]

import os import uuid from datetime import datetime from tensorflow.keras.callbacks import ModelCheckpoint # Generate a unique timestamp to be included in the checkpoint file name timestamp = datetime.now().strftime("%Y%m%d_%H%M%S") unique_id = str(uuid.uuid4())[:8] # Use the first 8 characters of a UUID as an identifier checkpoint_filepath = f"my_model_{timestamp}_{unique_id}.h5" # Check if the checkpoint file exists if os.path.exists(checkpoint_filepath): # Delete the existing file os.remove(checkpoint_filepath) # Create ModelCheckpoint callback to save only the best model based on validation accuracy model_checkpoint_callback = ModelCheckpoint( filepath=checkpoint_filepath, save_best_only=True, monitor='val_accuracy', mode='max', verbose=1 ) # Update the fit() function to use the model_checkpoint_callback history = avg_model.fit( train, epochs=50, verbose=1, callbacks=[model_checkpoint_callback], validation_data=val, validation_steps=16 )

here is i want to trained ensable learning model getting error.
ValueError: Unable to create dataset (name already exists)
i had tried many time but still error 

the list of models that is ensambled

vgg16_model = tf.keras.models.load_model('/kaggle/input/modelsforensable/best_model.h5')
vgg19_model = tf.keras.models.load_model('/kaggle/input/modelsforensable/vgg19_best_model.h5')
inceptionv3_model = tf.keras.models.load_model('/kaggle/input/modelsforensable/inception_best_model.h5')
resnet101v2_model = tf.keras.models.load_model('/kaggle/input/modelsforensable/resnet101v2_best_model.h5')

model1= Model(inputs = vgg16_model.inputs, outputs = vgg16_model.outputs)
model2= Model(inputs = vgg19_model.inputs, outputs = vgg19_model.outputs)
model3= Model(inputs = inceptionv3_model.inputs , outputs = inceptionv3_model.outputs)
model4= Model(inputs = resnet101v2_model.inputs, outputs = resnet101v2_model.outputs)

models = [ model1, model2, model3, model4]


# average ensemble model

# import Average layer
from tensorflow.keras.layers import Average, Dense, Dropout

input = Input(shape=(256, 256, 3))  # input layer

# get output for each input model
model_outputs = [model(input) for model in models]

# take average of the outputs
x = Average()(model_outputs)

x = Dense(100, activation='relu')(x)
x = Dropout(0.3)(x)
output = Dense(38, activation='softmax', name='output')(x) # output layer

# create average ensembled model
avg_model = Model(inputs = input, outputs = output)